Contamination MapTime series / Model
MantisV2
- Organisation
- Huawei Noah's Ark Lab (Paris)
- Family
- Mantis · version V2
- Released
- 2026-02-19 (arXiv v1 submission of the MantisV2 report, https://arxiv.org/abs/2602.17868v1 (HF repo created 2025-11-27 per HF API))
- Checkpoints
- paris-noah/MantisV2
- Weights
- Open · apache-2.0
- Links
- PaperModel cardCode
Training data disclosure: Full list
Level determined for 100% of benchmark cells.
- TestTest data seen0
- SeenSeen, split unclear0
- TrainTraining data seen0
- Related data seen0
- NoNot seen49
- ?Unknown0
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Other declared training data
-
CauKer 2M (synthetic, paris-noah/CauKer2M) · split: unknown
Purely synthetic (Gaussian-process kernel composition + structural causal models); released on HF.
Evidence
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
raw.githubusercontent.com · readme · Pre-training section · retrieved 2026-10-15 “Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
arxiv.org · paper · Section 3 · retrieved 2026-10-15 “In the remainder of the paper, we refer to the original Mantis architecture trained under this protocol as Mantis+.”
arxiv.org · paper · Section 3 · retrieved 2026-10-15
Benchmark cells
GIFT-Eval
Not seen Jena Weather (GIFT-Eval) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen BizITObs - Application · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen BizITObs - Service · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen BizITObs - L2C · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen Bitbrains - Fast Storage · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen Bitbrains - rnd · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen Restaurant (Recruit Restaurant Visitor Forecasting) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen ETT1 (GIFT-Eval) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen ETT2 (GIFT-Eval) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen Loop Seattle (LibCity) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen SZ-Taxi (LibCity) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen M_DENSE (LibCity) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen Solar (LSTNet, 137 Alabama PV plants) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen Hierarchical Sales (Mancuso et al., 2020) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen M4 Yearly · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen M4 Quarterly · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen M4 Monthly · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen M4 Weekly · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen M4 Daily · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen M4 Hourly · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen Hospital · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen COVID Deaths · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen US Births · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen Saugeen River Flow · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen Temperature Rain · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen KDD Cup 2018 · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen Car Parts · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen ElectricityLoadDiagrams20112014 (UCI) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Long-horizon forecasting (ETT, Electricity, Traffic, Weather, Exchange, ILI)
Not seen ETTh1 · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen ETTh2 · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen ETTm1 · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen ETTm2 · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen Electricity (ECL, 321 clients) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen Traffic (PeMS San Francisco Bay, 862 sensors) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen Weather (Jena, 2020; Autoformer) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen Exchange-Rate · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen ILI (CDC influenza-like illness, national) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Time-series anomaly detection
Not seen TSB-AD-U (TSB-AD univariate) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen TSB-AD-M (TSB-AD multivariate) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen UCR Time Series Anomaly Archive (Hexagon ML/UCR, KDD Cup 2021) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen Server Machine Dataset (SMD) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen Mars Science Laboratory (MSL) telemetry anomaly dataset · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen Soil Moisture Active Passive (SMAP) telemetry anomaly dataset · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen Pooled Server Metrics (PSM) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen Secure Water Treatment (SWaT) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen Yahoo S5 Labeled Anomaly Detection Dataset (Yahoo Webscope S5) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen Numenta Anomaly Benchmark (NAB) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Time-series classification
Not seen UCR Time Series Classification Archive (2018) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Not seen UEA Multivariate Time Series Classification Archive (2018) · absent from the complete list of training data
Evidence (2)
“We have open-sourced [CauKer 2M](https://huggingface.co/datasets/paris-noah/CauKer2M), the synthetic data set we used to pre-train the two versions of Mantis, resulting in [MantisPlus](https://huggingface.co/paris-noah/MantisPlus) and [MantisV2](https://huggingface.co/paris-noah/MantisV2) checkpoints.”
“Final pre-training. Based on these findings, we perform the final pre-training using 2 million synthetic time series generated by CauKer for 200 epochs.”
Notes
Refined architecture (kernel size 41, head dim 32, upgraded Transformer) trained on 2M CauKer synthetic series. Developers switched to synthetic data explicitly to avoid the UCR/UEA leakage of the original Mantis corpus, so no ts-classification data in pre-training.
Sources
- MantisV2 report (arXiv v1) · paper · retrieved 2026-10-15
- Mantis (arXiv v2, ICML 2026) · paper · retrieved 2026-10-15
- vfeofanov/mantis README · readme · retrieved 2026-10-15
- CauKer2M dataset card · dataset card · retrieved 2026-10-15